Design And Pricing Of Barrier Option Convertible Bond Based On LSTM Neural Network

Shanghai Composite Index of Chinese A-share has fallen 25% in year 2018.With financial risk releasing,convertible bond has become a beautiful scenery of risk prevention and control in securities market.Foreign convertible bond market is developing rapidly,and domestic convertible bond pricing research is relatively backward.Therefore,this paper first introduces the theory of convertible bond pricing and deep learning,and builds the model that spliting the value of convertible bond into option and corporate bond.In the B-S option pricing model,the measurement and fitting of volatility has always been the focus of academic research.In recent years,many studies have applied machine learning technology to financial markets.Advanced data analysis technology has achieved good results in financial markets,whether for financial data prediction or classification.LSTM neural network has advantages in processing time series data,and deep learning is also good at processing non-linear data.Therefore,with the above two advantages,we select the appropriate convertible bond data and compare the fitting of volatility between LSTM neural network and traditional linear GARCH model in empirical process to improve the option pricing of convertible bonds by using LSTM.In order to prevent and control financial risks,we need to carry out financial innovation,enrich financial markets and increase the ways for investors to hedge risks.So in the following part of this paper,we introduce the singular option convertible bond boldly.After introducing the related theories of singular options,this paper takes barrier convertible bonds as an example.A new obstacle convertible bond arbitrage model is developed.Based on the geometric Brownian motion of stock price,the LSTM neural network is used to fit the volatility.The Monte Carlo simulation of four obstacle convertible bonds is carried out.The pricing results are compared with those of ordinary convertible bonds,and the arbitrage method of obstacle convertible bonds is introduced.